Alex Susu via llvm-dev
2018-Jul-07 00:52 UTC
[llvm-dev] LoopVectorize fails to vectorize more complex loops
Hello.
Could you please tell me why the first loop of the following program (also
maybe the
commented loop) doesn't get vectorized with LoopVectorize (from a recent
LLVM build from
the SVN repository from Jun 2018)?
typedef short TYPE;
TYPE data[1400][1200];
void kernel_covariance(int m, int n, TYPE mean[1200]) {
int i, j, k;
for (j = 0; j < m; j++)
{
mean[j] = 0.0;
for (i = 0; i < n; i++)
mean[j] += data[j][i];
//mean[j] /= float_n;
}
// This loop gets vectorized
for (i = 0; i < n; i++)
for (j = 0; j < m; j++)
data[i][j] -= mean[j];
/*
// This loop doesn't get vectorized either:
for (i = 0; i < m; i++)
for (j = i; j < m; j++)
{
cov[i][j] = 0.0;
for (k = 0; k < n; k++)
cov[i][j] += data[i][k] * data[j][k];
sum += data[i][k] * data[j][k];
cov[i][j] /= (float_n - 1.0);
cov[j][i] = cov[i][j];
}
*/
}
For the first loop I get the following debug info from clang and opt:
LV: Checking a loop in "kernel_covariance" from test.c:10:7
LV: Loop hints: force=? width=0 unroll=0
LV: Found a loop: for.body3.us
LV: PHI is not a poly recurrence.
LV: PHI is not a poly recurrence.
LV: Found an unidentified PHI. %2 = phi i16 [ 0, %for.body.us ], [
%add.us,
%for.body3.us ], !dbg !48
LV: Can't vectorize the instructions or CFG
LV: Not vectorizing: Cannot prove legality.
Thank you,
Alex
Nema, Ashutosh via llvm-dev
2018-Jul-07 02:47 UTC
[llvm-dev] LoopVectorize fails to vectorize more complex loops
Hi Alex,
This seems like LICM problem, where it's not able to promote the memory
reference to scalar because of unknown aliasing.
This result recurrence store remain in the loop body and loop vectorization
fails to vectorize.
Loop Versioning LICM pass can help in this case.
- Ashutosh
-----Original Message-----
From: llvm-dev [mailto:llvm-dev-bounces at lists.llvm.org] On Behalf Of Alex
Susu via llvm-dev
Sent: Saturday, July 7, 2018 6:22 AM
To: llvm-dev <llvm-dev at lists.llvm.org>
Subject: [llvm-dev] LoopVectorize fails to vectorize more complex loops
Hello.
Could you please tell me why the first loop of the following program (also
maybe the commented loop) doesn't get vectorized with LoopVectorize (from a
recent LLVM build from the SVN repository from Jun 2018)?
typedef short TYPE;
TYPE data[1400][1200];
void kernel_covariance(int m, int n, TYPE mean[1200]) {
int i, j, k;
for (j = 0; j < m; j++)
{
mean[j] = 0.0;
for (i = 0; i < n; i++)
mean[j] += data[j][i];
//mean[j] /= float_n;
}
// This loop gets vectorized
for (i = 0; i < n; i++)
for (j = 0; j < m; j++)
data[i][j] -= mean[j];
/*
// This loop doesn't get vectorized either:
for (i = 0; i < m; i++)
for (j = i; j < m; j++)
{
cov[i][j] = 0.0;
for (k = 0; k < n; k++)
cov[i][j] += data[i][k] * data[j][k];
sum += data[i][k] * data[j][k];
cov[i][j] /= (float_n - 1.0);
cov[j][i] = cov[i][j];
}
*/
}
For the first loop I get the following debug info from clang and opt:
LV: Checking a loop in "kernel_covariance" from test.c:10:7
LV: Loop hints: force=? width=0 unroll=0
LV: Found a loop: for.body3.us
LV: PHI is not a poly recurrence.
LV: PHI is not a poly recurrence.
LV: Found an unidentified PHI. %2 = phi i16 [ 0, %for.body.us ], [
%add.us, %for.body3.us ], !dbg !48
LV: Can't vectorize the instructions or CFG
LV: Not vectorizing: Cannot prove legality.
Thank you,
Alex
_______________________________________________
LLVM Developers mailing list
llvm-dev at lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
Caballero, Diego via llvm-dev
2018-Jul-07 03:37 UTC
[llvm-dev] LoopVectorize fails to vectorize more complex loops
Agreed on the aliasing issue. You could also rewrite your loop like this:
for (j = 0; j < m; j++)
{
TYPE red = 0.0;
for (i = 0; i < n; i++)
red += data[j][i];
mean[j] = red;
}
Diego
> -----Original Message-----
> From: llvm-dev [mailto:llvm-dev-bounces at lists.llvm.org] On Behalf Of
> Nema, Ashutosh via llvm-dev
> Sent: Friday, July 6, 2018 7:48 PM
> To: Alex Susu <alex.susu at gmail.com>
> Cc: llvm-dev at lists.llvm.org
> Subject: Re: [llvm-dev] LoopVectorize fails to vectorize more complex loops
>
> Hi Alex,
>
> This seems like LICM problem, where it's not able to promote the memory
> reference to scalar because of unknown aliasing.
> This result recurrence store remain in the loop body and loop vectorization
> fails to vectorize.
> Loop Versioning LICM pass can help in this case.
>
> - Ashutosh
>
> -----Original Message-----
> From: llvm-dev [mailto:llvm-dev-bounces at lists.llvm.org] On Behalf Of
Alex
> Susu via llvm-dev
> Sent: Saturday, July 7, 2018 6:22 AM
> To: llvm-dev <llvm-dev at lists.llvm.org>
> Subject: [llvm-dev] LoopVectorize fails to vectorize more complex loops
>
> Hello.
> Could you please tell me why the first loop of the following program
(also
> maybe the commented loop) doesn't get vectorized with LoopVectorize
> (from a recent LLVM build from the SVN repository from Jun 2018)?
>
> typedef short TYPE;
> TYPE data[1400][1200];
>
> void kernel_covariance(int m, int n, TYPE mean[1200]) {
> int i, j, k;
>
> for (j = 0; j < m; j++)
> {
> mean[j] = 0.0;
> for (i = 0; i < n; i++)
> mean[j] += data[j][i];
> //mean[j] /= float_n;
> }
>
> // This loop gets vectorized
> for (i = 0; i < n; i++)
> for (j = 0; j < m; j++)
> data[i][j] -= mean[j];
>
> /*
> // This loop doesn't get vectorized either:
> for (i = 0; i < m; i++)
> for (j = i; j < m; j++)
> {
> cov[i][j] = 0.0;
> for (k = 0; k < n; k++)
> cov[i][j] += data[i][k] * data[j][k];
> sum += data[i][k] * data[j][k];
> cov[i][j] /= (float_n - 1.0);
> cov[j][i] = cov[i][j];
> }
> */
> }
>
> For the first loop I get the following debug info from clang and opt:
> LV: Checking a loop in "kernel_covariance" from test.c:10:7
> LV: Loop hints: force=? width=0 unroll=0
> LV: Found a loop: for.body3.us
> LV: PHI is not a poly recurrence.
> LV: PHI is not a poly recurrence.
> LV: Found an unidentified PHI. %2 = phi i16 [ 0, %for.body.us ], [
%add.us,
> %for.body3.us ], !dbg !48
> LV: Can't vectorize the instructions or CFG
> LV: Not vectorizing: Cannot prove legality.
>
> Thank you,
> Alex
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev